Generating a Contact Matrix for Aged Care Settings in Australia: an agent-based model study

TL;DR

This study addresses the need for high-resolution, context-specific contact data to inform infectious disease modelling in aged care. It develops an agent-based model (ABM) within a synthetic, single-story Australian facility to generate proximity-based contact matrices across care levels and staff shifts, incorporating detailed movement, task scheduling, and a Wells-Riley–based transmission module. The results show pronounced heterogeneity in contacts by resident care level and shift, with vaccination scenarios substantially reducing transmission, especially when both staff and residents are boosted. The work demonstrates the utility of ABMs for evaluating targeted infection control in enclosed, high-risk settings and provides a transferable framework for facility-specific outbreak preparedness.

Abstract

This study presents an agent-based model (ABM) developed to simulate staff and resident interactions within a synthetic aged care facility, capturing movement, task execution, and proximity-based contact events across three staff shifts and varying levels of resident care. Contacts were defined by spatial thresholds (1.5 m and 3 m) and cumulative duration, enabling the generation of detailed contact matrices. Simulation results showed that low and medium care residents experienced the highest frequency of interactions, particularly with staff on morning and afternoon shifts, while high care residents and night staff had substantially fewer contacts. Contact rates varied significantly by care level and shift, confirmed through Poisson-based regression modelling. Temporal analyses revealed clustering of high-risk contacts during structured daily routines, especially communal and care activities. An integrated airborne transmission module, seeded with a single infectious staff member, demonstrated that infection risk was highest during high-contact shifts and among medium care residents. Vaccination scenarios reduced predicted transmission by up to 68\%, with the greatest impact observed when both staff and residents were vaccinated. These findings highlight the importance of accounting for contact heterogeneity in aged care and demonstrate the utility of ABMs for evaluating targeted infection control strategies in high-risk, enclosed environments.

Figures (8)

• Figure 1: Proposed agent-based simulation pipeline used to model contact patterns and transmission scenarios in Australian aged care facilities.
• Figure 2: Task flow of the agents within the aged care home.
• Figure 3: Predicted contact matrix of different agent type categories (high care resident, low care resident, medium care resident, staff shift 1, staff shift 2 and staff shift 3) per person per day for contact radii of 1.5 metres (A) and 3 metres (B). The matrix was used a generalised linear model with a Poisson distribution to provide an adjusted estimation of contact rates.
• Figure 4: Time Series analysis of mean rolling contacts across agent types. Rolling sum of contacts between each agent type within the model calculated over 20-minute intervals with data aggregated from per-minute contact counts. Interactions are displayed for high care(a), medium care(b), and low care residents(c), as well as staff across three shifts: Shift 1(d), Shift 2(e), and Shift 3(f).
• Figure 5: Predicted contacts by time across a 24-hour simulation cycle, summarised in 20-minute intervals for residents (by care level) and staff (by shift). Each bubble represents the number of contacts in that interval; bubble size indicates the average contact duration in minutes.